Perceptions of Welfare in the European Union

Perceptions of Welfare in the European Union Background to Growing United: Upgrading Europe s Convergence Machine Jonathan Karver Gabriela Inchauste

Change in life satisfaction, 2005-2016 (%, annualized) Change in satisfaction w/national, 2005-2016 (%, annualized) Subjective measures of well-being are critical, not only in and of themselves, but also because perceptions often can lead to behaviors that either reinforce or undermine real economic trends. For instance, it is well recognized that consumer and investor confidence have real economic consequences, and as such most macroeconomic models include expectations as a key determinant of economic outcomes. Similarly, perceptions over individual well-being and inequality often can lead to political support for fiscal or other policy reforms or to strong opposition over reforms that are perceived to be unfair (World Bank, 2017; Levy and Walton, 2005). In extreme settings, perceptions of unfairness can lead to social uprising or conflict (World Bank, 2017). In the context of the European Union, perceptions of the economic situation have been persistently below what macroeconomic trends would suggest. Despite a positive (albeit weak) relationship between GDP per capita growth and life satisfaction (Figure 1A), the relationship between GPD per capita and satisfaction with the national economy is much less clear (Figure 1B). People seem to be unhappy about the present and pessimistic about the future. Some of this gloominess is justified. Although the early 2000s was a period of growth and convergence in incomes among EU countries, differences in levels of incomes across countries are still large. Moreover, convergence across countries was stalled following the global financial crisis, while within country inequality has increased for several countries in the region (Inchauste and Karver, 2018). The observed increases in GDP per capita reflect the fact that some households have experienced impressive gains over the decade, while for others this has not necessarily been the case (Gorka et al, 2017). However, individuals who are objectively better off also seem to be pessimistic. In fact, many Europeans say they are not doing better when they seem to be doing quite well using objective measures of well-being. Figure 1. GDP per capita growth and change in satisfaction A. GPD per capita and overall life satisfaction 6 B. GDP per capita and satisfaction with national economy 20 4 BGR ROU 2 EST LVA LTU HRV BEL AUT CZE DNK FINFRA GBR HUN SVK POL LUX DEU MLT PRT 0NLD IRL SVN SWE CYP -4-2 ITAESP 0 2 4 6-2 -4 GRC -6-8 GDP per capita growth, 2005-2015 (annualized) 15 DEU HRV MLT 10 PRT HUN POL 5 NLD SWECZE BGR ROU SVK BEL 0AUT LUX LVA -4-2 DNK GBR EST ITA 0 2 LTU 4 6 FRA -5 IRL CYP FIN SVN -10 GRC -15 ESP -20 GDP per capita growth, 2005-2015 (annualized) Source: Source: Eurobarometer & World Bank WDI 1

In this paper, we explore alternative hypotheses to explain why Europeans are overly pessimistic that can be understood in the context of individual welfare and the larger economic situation of each country. Our analysis is driven by two key questions: Why is there a mismatch between perceptions and realities of the economy? And what explains the persistent pessimism of Europeans about the economy? We complement these two underlying questions by exploring the individual characteristics associated with an over- or under-estimation of one s own welfare based on subjective and objective measures. We propose the following six key hypotheses that could explain the gloominess throughout the region: Why are Europeans over-pessimistic about their own welfare? 1. Wellbeing is relative. Europeans compare themselves with others, and observe that they are relatively worse off, even though they are doing much better than in the past. 2. Expectations are too high. European expectations are growing faster than their realities. 3. Uncertainty over the future. Europeans are worried about the future. Beyond the scope of individual circumstances, what factors explain Europeans over-pessimism about the economy? 4. They increasingly view rising inequality as unfair. 5. They lack trust in institutions. 6. Pessimism is inflated by populist surges. Eurobarometer The European Commission carries out a set of annual surveys on public opinion in all 28-member states collectively known as Eurobarometer, comprised of a Standard, repeated survey complemented by rotating surveys through the year that focus on specific issues of public opinion. The Standard Eurobarometer represents a yearly repeated cross-sectional sample of adults in Europe and covers opinions about welfare, the economic situation, and trust in institutions, among others. We make use of data between 2005 and 2015 and estimate regional and sub-regional averages for the European Union. Life in Transition Survey (LiTS) The Life in Transition Survey (LiTS), carried out by the European bank for Reconstruction and Development in three rounds (2006, 2010, 2016), captures attitudes and perceptions of welfare, inequality, and societal norms (among others) across a large sample of countries in the Europe and Central Asia (ECA) region (typically, EU17 countries have been included only as benchmarking countries, rotating on a round by round basis). Results from each round of the survey are meant to be nationally representative of the adult (18+) population of each country, providing valuable information about perceptions that is consistent across space and time. We focus primarily on results from the 2010 and 2016 rounds. 1 Most countries covered in the sample represent Central and Eastern Europe (CEE). 1 Countries evaluated in both years from the EU: Bulgaria, Croatia, Czech Rep., Estonia, Hungary, Germany, Italy, Latvia, Lithuania, Poland, Romania, Slovak Rep., and Slovenia. Cyprus, and Greece were added in 2016, whereas France, the UK, and Sweden had each been surveyed in 2010 but no follow-up in 2016. 2

Greece Bulgaria Romania Hungary Portugal Italy Lithuania Croatia Slovakia Latvia Spain Cyprus Poland Estonia France Czech Rep. Austria Germany Slovenia Belgium Malta Ireland Great Britain Finland Netherlands Sweden Luxembourg Denmark Perception of national economy Annualized growth rate Europeans persistently believe the economic situation in their countries is not good. Perceptions regarding the current situation of the national economy are particularly dire, especially when compared to overall life satisfaction in the EU. Less than 4 in the EU-15 and just one third in CEE believe the economic situation is positive (Figure 2 and Figure 3). Furthermore, the expectations among most EU residents is that the national economy will not improve in the near future (Figure 4), even when the economies are actually doing (slightly) better: economic growth since 1990 has been strong despite the economic cycle, and despite pre-crisis/post-crisis differences (though in past two years most countries have observed a gradual return to pre-crisis trends). Figure 2. Perceptions on the current state of affairs with respect to the national economy, 2005-2016 EU-15 CEE EU-28 10 10 10 8 8 8 6 6 6 4 4 4 2 2 2 2005 2010 2016 2005 2010 2016 2005 2010 2016 Source: Eurobarometer Figure 3. Perceptions on the current state of affairs with respect to the national economy by country, 2016 10 9 8 7 6 5 4 3 2 1 5.0 4.0 3.0 2.0 1.0 0.0-1.0-2.0-3.0 Source: Eurobarometer & World Bank WDI. Very good or rather good Annualized GDP growth 2005-2015 3

10 8 6 4 2 Figure 4. Expectations for the national economy in the next 12 months, 2005-2016 EU-15 CEE EU-28 2005 2008 2010 2016 10 8 6 4 2 2005 2008 2010 2016 10 8 6 4 2 2005 2008 2010 2016 Source: Eurobarometer Most Europeans perceive very little improvement in their well-being relative to the past and are pessimistic about the future. Perceptions of mobility in the EU (among mostly CEE countries) are generally low, with the vast majority of respondents reporting that their welfare has remained unchanged relative to 4 years prior (Figure 5). In addition, they expect no improvement in the future (4 years later). Pessimism with respect to individual welfare, while not unique to the EU, is substantially worse than in other non-eu ECA countries. Latvia Estonia Poland Lithuania Romania Germany Czech Rep. Bulgaria Hungary Slovak Rep. Croatia Italy Slovenia Cyprus Greece Figure 5. Perceived individual welfare relative to the past and future, by country, 2016 EU Non-EU 2 4 6 8 10 Uzbekistan Kyrgyz Rep. Tajikistan Mongolia Kosovo Moldova Georgia Kazakhstan Azerbaijan Turkey Montenegro Albania Serbia FYR Macedonia Bosnia and Herz. Armenia Belarus Russia Ukraine 2 4 6 8 10 Source: LiTS III (2016). Note: past and future both 4 years from present. 4

Part of that pessimism is explained by individual characteristics. Older and less educated adults display a lower probability of reporting improvements in welfare relative to the past (Figure 6). Older individuals are less likely to report continuous improvement from past to future (welfare in future expected to be higher than present, welfare today higher than past). Perceived mobility is highest among those in the 25-34 and 35-44 age groups and those with tertiary education, and worst among those 55 and older. The unemployed are significantly less likely to report improvements in welfare relative to the past, but they are also significantly more hopeful when looking at the future. However, even wealthier individuals are pessimistic. Surprisingly, wealthier respondents are not significantly more likely to perceive that their welfare has improved relative to the past. Moreover, relatively wealthy individuals (those in the 7 th decile) are significantly less likely to perceive that their welfare will improve in the future. More generally, wealthier individuals are not significantly more likely to perceive their welfare will continuously improve. Individuals that see no improvement in their well-being are more likely to say they are dissatisfied with economy. At the individual level, perceived welfare is highly correlated with views on the national economy, and this finding is consistent across various measures of individual welfare. In general, males and those with higher levels of education are more likely to be satisfied with the national economy. Individuals that see no improvement in their well-being are more likely to say they are dissatisfied with economy (Figure 7A). In general, perceptions about own welfare influence perceptions about the economy: the probability of dissatisfaction is significantly lower among those reporting high life satisfaction and those who consider themselves in the middle class or higher (Figure 7B). 5

Marginal effect in percentage points Round decile=2 decile=3 decile=4 decile=5 decile=6 decile=7 decile=8 decile=9 decile=10 HH head Male Age: 25-34 Age: 35-44 Age: 45-54 Age: 55-64 Age: 65+ Education: Education: Tertiary Unemployed Urban Home ownership Marginal effect in percentage points Round decile=2 decile=3 decile=4 decile=5 decile=6 decile=7 decile=8 decile=9 decile=10 HH head Male Age: 25-34 Age: 35-44 Age: 45-54 Age: 55-64 Age: 65+ Education: Education: Tertiary Unemployed Urban Home ownership Marginal effect in percentage points Round decile=2 decile=3 decile=4 decile=5 decile=6 decile=7 decile=8 decile=9 decile=10 HH head Male Age: 25-34 Age: 35-44 Age: 45-54 Age: 55-64 Age: 65+ Education: Secondary Education: Tertiary Unemployed Urban Home ownership Figure 6. Correlates of improvement in welfare relative to the past, future, and continuous (a) Improvement relative to past 1 5% -5% -1-15% -2-25% 1-5% 5% -1-15% -2-25% -3-35% -4 (b) Improvement expected in future 4% 2% -2% -4% -6% -8% -1-12% -14% (c) Continuous improvement Source: Life in Transition Survey (LiTS), 2010-2016. Marginal effects estimated from Probit regression with additional controls (See full regression results in appendix). indicates p<0.05. Sample is limited to those with the WFA defined. 6

Marginal effect in percentage points Marginal effect in percentage points 15% 1 5% Figure 7. Correlates of dissatisfaction with the national economy A. 2010-2016 B. 2005-2016 -5% Improvement Improvement No in future only relative to improvement past only No decile controls Decile controls Source: Left: Life in Transition Survey (LiTS). Right: Eurobarometer. Marginal effects estimated from Probit regression with additional controls (See full regression results in appendix). indicates p<0.05. 15% 1 5% -5% -1-15% -2 Overall Perceive satisfaction that quality w/life of life was better in past The lower middle class of society The middle class of society The upper middle class of society The higher class of society Relative Welfare People compare themselves to others that may be doing better. People care about their wellbeing compared to others, even if their wellbeing has improved in absolute terms. Most Europeans (in a sample of mostly CEE countries, Figure 8) compare themselves to friends and neighbors or Western Europeans, highlighting perceived inequalities in the region. Individuals are more likely to compare themselves to peers than to the past if they believe mobility will happen in the future. Figure 8. Comparison group for current economic situation, 2016 Bulgaria Croatia Cyprus Czech Rep. Estonia Germany Greece Hungary Italy Latvia Lithuania Poland Romania Slovak Rep. Slovenia 2 4 6 8 10 Friends/neighbors Parents W. Europe Domestic elite Pre 1989/1991 Neighboring countries (ex-communist) Don't know Source: Life in Transition Survey (LiTS), 2016 7

Marginal effect in percentage points increase in probability of benchmarking against peers/w. Europe Mobility in future only Mobility relative to past only No mobility Age: 25-34 Age: 35-44 Age: 45-54 Age: 55-64 Age: 65+ Education: Secondary Education: Tertiary decile=2 decile=3 decile=4 decile=5 decile=6 decile=7 decile=8 decile=9 decile=10 Younger generations in CEE countries compare themselves to Western Europe more than do older generations (Figure 9). The age group that compares itself most with Western Europe (45-54 years of age at the time of the survey) is the group that was young (18-30) during the time of transition. For most of the CEE countries, this occurred in the early 1990s. Wealthier and more educated individuals are more likely to compare themselves to Western Europe than to their own parents. Individuals with tertiary education are 16 percentage points more likely to compare themselves with Western Europe than with their own peer (Figure 10). Similarly, those who are in the top 30 percent of the distribution are over 15 percentage points more likely to compare themselves with Figure 9. Comparison group for current economic situation by age group among mostly CEE countries (2016) 65+ 55-64 45-54 35-44 25-34 18-24 5 10 Past (parents/pre-transition) Present (friends/neighbors/domestic elite/neighboring excommunist) Future (W. Europe) Source: Life in Transition Survey (LiTS), 2016. Countries included in average: Bulgaria, Croatia, Czech Rep., Estonia, Hungary, Germany, Italy, Latvia, Lithuania, Poland, Romania, Slovak Rep., Slovenia, Cyprus, and Greece Western Europe than with their peers. This could explain why even wealthier individuals are gloomy, as it seems that they are still feeling relatively worse off, despite the fact that they are doing quite well in absolute terms. Figure 10. Correlates of benchmarking welfare, peers vs. Western Europe 3 25% 2 15% 1 5% -5% -1-15% Peers W. Europe Source: Life in Transition Survey (LiTS), 2016. Marginal effects estimated from Probit regression with additional controls (See full regression results in appendix). indicates p<0.05 8

Marginal effect in percentage points Expectations People compare themselves to where they think they should be. The observed gloominess in European s perceptions could also reflect disappointment over slower than expected gains in welfare. As shown above, this could reflect that they are comparing themselves to higher income countries in Europe, with expectations likely growing faster than real incomes. Moreover, many Europeans are not convinced they are doing better than their parents, nor do they expect future generations to be better off. Among a sample of respondents from mostly CEE countries, perceived mobility in terms of benchmarking against past and future generations reveals a pessimistic trend. On average, less than 4 believe they are better off than their parents and less than 4 believe the next generation will be better off (Figure 11). Lithuania Estonia Czech Rep. Latvia Romania Slovenia Slovak Rep. Croatia Average Poland Hungary Germany Italy Bulgaria Figure 11. Level of agreement, 2016 Better off relative to parents Future generation will be better off 2 4 6 8 10 Lithuania Poland Latvia Estonia Germany Croatia Bulgaria Average Hungary Romania Slovak Rep. Czech Rep. Italy Slovenia 2 4 6 8 10 Strongly agree Agree Neither disagree nor agree Disagree Strongly disagree Source: LiTS III (2016). Evidence suggests that intergenerational mobility is higher than perceptions would suggest. Nearly one third of respondent in select EU countries have a higher level of education than one or both of their parents. Yet, only 46 percent of respondents experiencing educational intergenerational mobility believe they are better off than their parents, compared to 38 percent of those without mobility (these differences are Figure 12. Observed mobility and perceptions of improvement in welfare 2 15% 1 5% -5% Not controlling for decile decile Source: LiTS III (2016). Marginal effects estimated from Probit regression with additional controls (See full regression results in appendix). indicates p<0.05 Controlling for Better off than parents No improvement in welfare Continuous improvement in welfare 9

Marginal effect in percentage points increase probability of perceiving improvement relative to past/future generations Current (perceived) welfare ladder Mobility in future only Mobility relative to past only No mobility Age: 25-34 Age: 35-44 Age: 45-54 Age: 55-64 Age: 65+ Education: Secondary Education: Tertiary Unemployed Urban Home ownership statistically significant). Despite this disconnect, the correlation between perceived and observed mobility relative to their parents is high those who have a higher education than parents are 7 percentage points more likely to perceive that they are better off (though not more likely to perceive mobility in the short/medium term overall). Observed mobility (having more education than parents) is associated with an 8-15 percentage point increase in the probability of reporting improvement vis-à-vis parents (Figure 12). Although individual characteristics color future expectations, even those who claim to have experienced upward mobility are pessimistic about the future. Older individuals recognize that they are better off than their parents, as do those owning their home (Figure 13). In contrast, those who are unemployed don t believe they are better off than their parents, and are pessimistic about the opportunities available to future generations. More generally, over-pessimistic people are more likely to view themselves worse off than their parents and view the next generation worse off than them. What is surprising is that individuals with secondary and tertiary education and those who perceive they have had social mobility in the past are nevertheless pessimistic about the opportunities for future generations. Figure 13. Correlates of perceptions regarding expectations relative to past and future generations 3 25% 2 15% 1 5% -5% -1-15% -2 Better off than parents Next generation will be better off Better off than parents & same for next generation Source: Own estimates based on LiTS III (2016). Marginal effects estimated from Probit regression with additional controls (See full regression results in appendix). indicates p<0.05 10

2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 Uncertainty People may face increased uncertainty and could therefore be more pessimistic about the future. To the extent that jobs and incomes become more volatile, people may perceive increased uncertainty about a future for which they are not insured. Incomes and employment have indeed become more volatile over time (Figure 13). Higher shares of adults are working part-time or have a secondary job (Figure 14). EU-SILC data suggests that middle aged and those with tertiary education report higher shares of secondary employment. Real wages have stagnated in Europe since the crisis, as has the quality of employment. includes sample of OECD countries in the EU. Figure 14. Change in share of temporary contract employment over time, by age group EU-15 CEE 45% 4 35% 3 25% 2 15% 1 5% Figure 13. Change in real wages (2015 USD) over time, EU OECD countries 50,000 40,000 30,000 20,000 10,000 0 Source: OECD. Unweighted regional averages. Only 45% 4 35% 3 25% 2 15% 1 5% CEE EU17 15-24 25-54 55-64 15-24 25-54 55-64 Source: Eurostat. Regional average are unweighted. However, the evidence on the link between uncertainty about the future and the quality and outlook of employment is mixed. As mentioned earlier, unemployment is positively associated with future improvement in perceived welfare (perhaps capturing a measure of hope). However, no relationship exists between the quality of employment (hours worked, type of contract) and perceived improvement of welfare in future. A positive (albeit weak) relationship exists between the type of employment (selfemployed, employee, high skill/low skill) and expectations for the next 12 months (Figure 15). Evaluating these correlations among young adults only (those most afflicted by temporary contracts) or specific groups of countries (those with high shares of temporary employment) do not provide significant results. What does seem to matter more is age. The probability of negative expectations for the future increases nearly monotonically along the age distribution: that is, feelings of uncertainty with regards to the national 11

economy, household financial situation, national employment situation, and personal job situation are all more likely among older groups of individuals. However, supplementary country-level surveys provide evidence supporting some of the pessimism with regards to the future. Evidence from Poland seems to coincide with current labor market realities in the EU (CBOS 2017). The fear of loss of employment increased substantially after the crisis, and has only began to move slowly towards pre-crisis perceptions. It is lowest among those working in agriculture and those earning high incomes, and highest among employees dissatisfied with their financial situation, as well as among key occupational groups (e.g., service workers and unskilled workers). Individuals in low skilled occupations tend to have a less favorable outlook on the future. Figure 15. Correlates of negative outlook on economy, financial situation, employment 5% Employee: 4% low skill 3% Selfemployed: 2% high skill 1% Selfemployed: low skill -1% -2% Economy HH financial National Personal job situation employment situation Source: Eurobarometer 2005-2016. Marginal effects estimated from Probit regression with additional controls (See full regression results in appendix). indicates p<0.05 Perceptions of the economy (in its broadest sense) go hand in hand with perceptions of individual welfare: relative welfare, for example, is based on the premise of winners and losers, driven by national, regional, or even global trends in the economy. The same can be said of expectations, which are a function of the (perceived) outlook on aggregate improvements. The perception that others are doing better highlights concerns of unfairness and the ability (and willingness) of public institutions to confront it. As such, perceptions about the economic situation are an extension of perceptions of individual well-being in that they reflect a larger concern with aggregate welfare. Unfairness People are concerned about inequality, particularly if viewed as unfair. Most Europeans (among a sample of mostly CEE countries) believe inequality should be reduced (Figure 16). Although many Europeans believe individual effort should be rewarded (Figure 17), social injustice is perceived to be the most important reason for individuals to be in need (Figure 18). People who are more concerned about inequality tend to be more pessimistic about the economy (Figure 19). In general, perceptions of inequality are consistent with realities. These findings are reflective of a regional trend of increasing inequality within countries (despite convergence across), a significant share of which can be explained by inequality of opportunity. Perceptions of inequality suggest that most respondents (mostly in CEE) agree 12

Marginal effect in percentage points Frequency that inequality should be reduced (80 percent on average agree); however, an important (and growing) share of respondents believe larger income differences are needed to incentivize effort and that individual effort should be rewarded (Figure 17). Figure 16. Inequality gap should be reduced (2016) Slovenia Lithuania Latvia Estonia Romania Italy Croatia Hungary Slovak Rep. Bulgaria Average Germany Poland Czech Rep. 2 4 6 8 10 Strongly agree Agree Neither Disagree Strongly disagree Source: Life in Transition Survey (LiTS), 2010 and 2016. Figure 17. Spectrum of optimal inequality 0.25 0.20 0.15 0.10 0.05 0.00 1 2 3 4 5 6 7 8 9 10 Incomes should be more equal -----> Need larger differences to incentivize effort 2010 2016 Figure 18. Most important reason why people are in need, 2010-2016 10 8 6 4 2 21. 17.9% 43.6% 23.7% 35.1% 23.6% 11.7% 16.1% 2010 2016 It is an inevitable part of modern life Because of injustice in our society Because of laziness and lack of willpower Because they have been unlucky Figure 19. Dissatisfaction with the economy and perceptions of inequality 12% 1 8% 6% 4% 2% Belief that inequality should be reduced People in need because of social injustice Source: Life in Transition Survey (LiTS), 2010 and 2016. Right: Marginal effects estimated from Probit regression with additional controls (See full regression results in appendix). indicates p<0.05 13

Low trust in institutions Pessimism is associated with low trust in governments and EU institutions. Most people do not trust the political party system, national governments and, to a lesser extent, EU institutions (Figure 20). In many countries, trust is higher for supranational governments (EU commission) than national governments. Trust in the EU is substantially higher than trust in the national government, although it has decreased after reaching a high in 2008. Moreover, this trust gap is largest among CEE countries. Most EU respondents trust neither their own government nor the EU. These shares have worsened since 2008. The share of individuals across countries who trust their national government, the EU government, both or neither shows interesting trends: less than 2 of respondents in the EU trust both, and 29% of CEE respondents trust only the EU. 54% of EU respondents trust neither (Figure 21). 10 8 6 4 2 10 8 6 4 2 10 8 6 4 2 Source: Eurobarometer. Figure 20. Trust in institutions A. Trust in the political party system EU-15 CEE EU-28 10 10 8 8 6 6 4 4 2005 2008 2010 2016 2 2005 2008 2010 2016 2 B. Trust in the national government EU-15 CEE EU-28 10 10 8 8 6 6 2005 2008 2010 2016 4 2 2005 2008 2010 2016 4 2 C. Trust in the European Union EU-15 CEE EU-28 10 10 8 8 6 6 2005 2008 2010 2016 4 2 2005 2008 2010 2016 4 2 Don't Know Tend Not To Trust Tend To Trust 2005 2008 2010 2016 2005 2008 2010 2016 2005 2008 2010 2016 14

Marginal effect in percentage points Figure 21. Distribution of trust Malta Estonia Luxembourg Netherlands Finland Denmark Sweden Ireland Austria Portugal Belgium Bulgaria Germany Slovakia Lithuania Latvia Great Britain Romania Cyprus Hungary Czech Italy Poland Croatia Spain Slovenia France Greece Trust neither Trust national only Trust EU only Trust both Source: Eurobarometer 2 4 6 8 10 Individuals who do not trust political institutions tend to be more pessimistic about the economy and their own welfare. Individuals who trust both national and EU institutions are over 30 percentage points less likely to be dissatisfied with the economy (Figure 22) and 17 percentage points less likely to be dissatisfied with their own welfare. Figure 22. Dissatisfaction with economy/own welfare and trust in institutions -5% -1-15% -2-25% -3-35% Trust national only Trust EU only Trust both Economic situation bad/rather bad Source: Eurobarometer. Different colored bars represent coefficients from different estimations, where the outcome is either dissatisfaction with economy or life situation. Marginal effects estimated from Probit regression with additional controls (See full regression results in appendix). indicates p<0.05 15 Life situation overall bad/rather bad

Marginal effect in percentage points Marginal effect in percentage points Populism Populist movements exploit individuals pessimism and frustration, widening the gap between realities and perceptions. All of the previous hypotheses point to some rational explanation for the overpessimism of Europeans. Are populist movements exploiting this over-pessimism, making people believe that things are worse than they really are for political gain? There is some evidence this is the case. EU countries where (authoritarian) populism is relatively high (as measured by vote shares going to authoritarian populist parties using the Timbro index) are more likely to report the economic situation being bad/rather bad. More specifically, respondents in countries with higher shares of votes going to populist parties are more likely to be dissatisfied with the economic situation (Figure 23). Recent shifts in the political spectrum towards populism are also associated with perceived improvements relative to past (Figure 24), but are not statistically related to expectations about the future. No association between individual voting behavior captured in the LiTS for the most recent local, parliamentary and presidential elections, and measures of perceived welfare improvement relative to the past or future exists. Figure 23. Role of populism in perceptions of welfare 3 25% 2 15% 1 5% Perceive that quality of life was better in past Economic situation bad/rather bad Figure 24. Voting behavior and welfare improvement 15% 1 5% -5% -1 Voted for candidate/party program and promises in any elections Source: Eurobarometer (left) and LiTS (right), and the Timbro Index (https://timbro.se/allmant/timbroauthoritarian-populism-index-2016). High support for authoritarian populism is defined at the country level where the share of votes going to authoritarian populist parties in the most recent election (2016) exceeds 2 of the total (Timbro index 2016). Voted for a candidate-party program and promises in any elections refers to individuals who claim to have voted in the most recent local, parliamentary, or presidential elections based on candidate/party program and promises rather than the political affiliation, perceived competence/reputation of the candidate, or religion/ethnicity of the candidate/party (defined in the LiTS). Marginal effects estimated from Probit regression with additional controls (See full regression results in appendix). indicates p<0.05 Improvement relative to past Improvement expected in future High/very high support for populism Continuous improvement No improvement Our analysis thus far suggests that perceptions of own welfare and the economic situation in general are gloomier than the macroeconomic trends would suggest. A more in-depth analysis of this disconnect between perceptions and reality could uncover the specific circumstances that drive this wedge between 16

% correctly predicting welfare (+/- 1) perceptions and reality. The LiTS captures a measure of objective welfare through levels of consumption based on a set of expenditures; by comparing perceived and observed locations in the welfare distribution, we can get a better idea of particular groups of people that are feeling this surge in pessimism. Disconnect between observed and perceived levels of welfare In general, people are far off in identifying their location in the welfare distribution. Figure 25 below reports the distribution of individuals reporting their position on each step of a Cantril s ladder 2, by decile of the consumption distribution. While the extremes in perceptions line up best with the extremes of actual welfare, it is obvious there is a strong disconnect between observed and perceived welfare. In fact, only 12 percent of respondents in 2016 correctly identify their decile (based on their perceived location in the distribution), 36 percent believe they are further up the ladder, and the remaining 52 percent underpredict where they actually lie in the distribution. The situation improves if we consider a 1 step margin of error (to 31% correct prediction), but this is concentrated in the middle of the distribution and the disconnect remains substantial in the extremes. For example, less than 2 of individuals in the first decile actually believe they are in the in bottom three deciles and less than 15% of individuals in the top decile actually think they are in the top three deciles. Not surprisingly, correct welfare prediction is highest closest to the median (Figure 26). Figure 25. Perceived vs observed location on the welfare latter, 2016 10 9 8 7 6 5 4 3 2 1 1 2 3 4 5 6 7 8 9 10 Decile of the consumption distribution Step 10 Step 9 Step 8 Step 7 Step 6 Step 5 Step 4 Step 3 Step 2 Figure 26. Correct welfare prediction by decile 8 7 6 5 4 3 2 1 Step 1 1 2 3 4 5 6 7 8 9 10 Decile of consumption distribution Figure 27. Distribution of population above or below the observed and perceived median Source: LiTS 2010 and 2016. The observed decile is based on a consumption-based welfare aggregate including food, services, transportation, education, health, clothing, and durables and is only defined where all consumption groupings are non-missing. 10 9 8 7 6 5 4 3 2 1 2010 2016 Above observed and above perceived median Above observed and below perceived median Below observed and above perceived median Below observed and perceived median 2 The LiTS survey captures perceived welfare through a Cantril s ladder, which can be equated to deciles on the consumption (expenditure) distribution. 17

More broadly, 44 percent of individuals believe they are on the opposite side of the welfare distribution than where they actually fall, and among these, two thirds think they fall below it. Figure 27 simplifies the interpretation of the welfare mismatch by grouping individuals into one of four groups: (i) below the observed welfare median (1 st -5 th deciles of the consumption distribution) and below the perceived welfare distribution median (steps 1-5 in Cantril s ladder); (ii) below the observed but above the perceived median; (iii) above the observed but below the perceived median; (iv) and above the observed and perceived median. The most notable mismatch is among those in the top of the welfare distribution who think they are in the bottom (29% of all respondents, or two thirds of the top half of the welfare distribution). This is indicative of a mismatch that is more prominent among the relatively better off in each country. Correlates of under-estimation of own welfare Under-prediction of own welfare is strongly associated with age, level of education, and unemployment, whether we consider steps (deciles) away from one s actual welfare location within the country or the perception of being in the bottom half of the welfare distribution despite pertaining to the top half. Figure 28 reports the marginal effects associated with under-predicting one s location in the welfare distribution. Older respondents (up to 44 years of age) are between 3 and 12 percentage points less likely to perceive being in a lower position in the welfare distribution than what is observed relative to the youngest respondents in the sample (18-24 years of age). Meanwhile, under-prediction is significantly higher among the more educated respondents: individuals with secondary and tertiary education are 4 and 7 percentage points more likely to under-predict their welfare (relative to those with primary education or less). The picture changes slightly when we only consider the top half of the distribution and evaluate whether or not they feel they belong to the bottom half. Here, the more educated are between 15 and 33 percentage points less likely to perceive belonging to the bottom half of the welfare distribution, perhaps pointing to the increased vulnerability felt by the less educated among the wealthier half of respondents. 18

Round HH head Male Age: 25-34 Age: 35-44 Age: 45-54 Age: 55-64 Age: 65+ Education: Secondary Education: Tertiary Unemployed Urban Home ownership Round HH head Male Age: 25-34 Age: 35-44 Age: 45-54 Age: 55-64 Age: 65+ Education: Secondary Education: Tertiary Unemployed Urban Home ownership Marginal effect in percentage points Marginal effect in percentage points Figure 28. Correlates of under-prediction of own welfare (margin of error=1 step/decile) Figure 29. Correlates of perceiving being below the welfare median despite being above it 2 2 15% 1 1 5% -5% -1-2 -1-3 -15% -4 Source: LiTS 2010 and 2016. Marginal effects estimated from Probit regression with additional controls (See full regression results in appendix). indicates p<0.05 The evidence presented in this paper suggests that, while European s are more pessimistic than they should be given aggregate economic trends, there are a number of factors behind this sentiment that justify plenty of the gloom. For example, despite within country inequality remaining fairly flat in most countries, the convergence across countries observed prior to the crisis has stalled, and Europeans are aware of what is possible, especially in such an interdependent region as the EU. Europeans compare themselves with others, and observe that they are relatively worse off, even though they are doing much better than in the past. They increasingly view rising inequality as unfair. This is further exacerbated by uncertainty in the future and the vulnerability to shocks that still exists in most of the region, not to mention low trust in institutions that are meant to safeguard against these and national dialogues that feed off this pessimism. More research is needed to explore what the consequences of this pessimism are (for example, it s effect on populist political movements), but the analysis in this paper suggests that people feel more vulnerable to losses in welfare than hard data would suggest. 19

Gorka, S., W. Hardy, R. Keister, and P. Lewandowski. 2017. Tasks and skills in European labor markets. Background paper for Growing United, IBS Research Report 03/2017. Warsaw: Institute for Structural Research. Inchauste, G., and J. Karver. 2018. Understanding changes in inequality in the European Union. Background paper for Growing United. World Bank, Washington, D.C. Levy, S., and M. Walton. 2005. No Growth without Equity: Inequality, Interests and Competition in Mexico. London: Palgrave Macmillan. World Bank Group. 2017. World Development Report 2017: Governance and the Law. Washington, DC: World Bank. World Bank. https://openknowledge.worldbank.org/handle/10986/25880 20

Table A1. Correlates of improvement in welfare relative to the past, future, and continuous Improvement relative to past 21 Improvement expected in future Continuous improvement Round 0.007 0.065 0.022 [0.015] [0.017] [0.007] decile=2 0.012-0.002 0.022 [0.022] [0.027] [0.011] decile=3-0.026-0.018 0.004 [0.023] [0.025] [0.011] decile=4-0.023-0.017 0.004 [0.021] [0.024] [0.009] decile=5-0.007-0.025 0.008 [0.022] [0.026] [0.011] decile=6-0.019-0.035-0.002 [0.023] [0.026] [0.010] decile=7-0.020-0.063-0.010 [0.023] [0.024] [0.010] decile=8-0.017-0.052-0.002 [0.022] [0.028] [0.009] decile=9 0.020-0.039 0.010 [0.024] [0.025] [0.011] decile=10 0.041-0.035 0.022 [0.023] [0.024] [0.011] HH head -0.025 0.042 0.003 [0.011] [0.012] [0.006] Male 0.008-0.013 0.004 [0.010] [0.010] [0.005] Age: 25-34 0.030-0.019 0.009 [0.019] [0.022] [0.008] Age: 35-44 -0.000-0.108-0.017 [0.019] [0.023] [0.009] Age: 45-54 -0.052-0.189-0.052 [0.020] [0.022] [0.010] Age: 55-64 -0.121-0.271-0.085 [0.020] [0.023] [0.011] Age: 65+ -0.206-0.354-0.125 [0.021] [0.026] [0.011] Education: Secondary 0.019 0.008 0.008 [0.012] [0.014] [0.006] Education: Tertiary 0.075 0.043 0.029 [0.014] [0.020] [0.008] Unemployed -0.085 0.049-0.013 [0.019] [0.018] [0.009]

Urban 0.005 0.018 0.011 [0.012] [0.013] [0.006] Home ownership 0.023-0.037 0.001 [0.012] [0.014] [0.006] Observations 25757 25757 25757 Source: LiTS. Note: Marginal effects from Probit estimation (evaluated at means). Country fixed effects not shown. Robust standard errors in parenthesis. p<0.01, p<0.05, p<0.1 Table A2. Correlates of dissatisfaction with the national economy Individual welfare measure (I) (II) Overall satisfaction w/life -0.179 [0.005] Perceive that quality of life was better in past 0.120 [0.008] The lower middle class of society -0.020 [0.011] The middle class of society -0.124 [0.011] The upper middle class of society -0.166 [0.017] The higher class of society -0.134 [0.043] year==2010 0.148 [0.005] year==2016-0.063 [0.004] Male -0.067-0.086 [0.004] [0.008] Age: 25-34 0.000 0.005 [0.008] [0.018] Age: 35-44 0.011-0.023 [0.008] [0.019] Age: 45-54 0.016-0.012 [0.008] [0.019] Age: 55-64 -0.000-0.015 [0.009] [0.020] Age: 65+ -0.021-0.078 [0.010] [0.023] Secondary education or more -0.045-0.039 [0.004] [0.009] Student -0.026-0.080 [0.009] [0.021] Unemployed 0.102 0.118 22

[0.008] [0.017] Retired/Disabled 0.023 0.045 [0.007] [0.015] Self-employed 0.008-0.018 [0.007] [0.015] Household size 0.002-0.002 [0.001] [0.003] EU-13 0.257 0.163 [0.045] [0.076] Observations 68762 22123 Source: Eurobarometer. Note: Marginal effects from Probit estimation (evaluated at means). Country fixed effects not shown. Robust standard errors in parenthesis. p<0.01, p<0.05, p<0.1 (I) and (II) separate estimations using different measures of life satisfaction. Table A3. Correlates of benchmarking welfare, peers vs. Western Europe Independent variable Multinomial logit (base: past), no decile controls Multinomial logit (base: past), decile controls Peers W. Europe Peers W. Europe Mobility in future only 0.050-0.014 0.033 0.005 [0.046] [0.047] [0.051] [0.043] Mobility relative to past only -0.012-0.048-0.028 0.000 [0.044] [0.048] [0.052] [0.044] No mobility 0.020-0.060 0.015-0.041 [0.038] [0.039] [0.043] [0.036] HH head -0.009 0.023 0.004 0.000 [0.027] [0.026] [0.028] [0.025] Male 0.037-0.022 0.011-0.001 [0.026] [0.030] [0.026] [0.027] Age: 25-34 0.052-0.048 0.090-0.102 [0.038] [0.033] [0.044] [0.042] Age: 35-44 -0.029-0.004 0.047-0.092 [0.041] [0.036] [0.043] [0.042] Age: 45-54 -0.063 0.075 0.018-0.048 [0.043] [0.044] [0.044] [0.045] Age: 55-64 -0.051-0.019 0.007-0.100 [0.042] [0.039] [0.051] [0.040] Age: 65+ 0.015-0.080 0.037-0.121 [0.062] [0.062] [0.063] [0.060] Education: Secondary -0.068 0.109-0.099 0.101 [0.029] [0.034] [0.032] [0.033] Education: Tertiary -0.063 0.126-0.127 0.163 23

[0.034] [0.035] [0.040] [0.031] Unemployed -0.021-0.070-0.039-0.064 [0.037] [0.037] [0.040] [0.041] Urban 0.029 0.015 0.022 0.025 [0.024] [0.025] [0.032] [0.030] Home ownership -0.031-0.012-0.058 0.034 [0.034] [0.039] [0.038] [0.038] decile=2 0.054 0.066 [0.051] [0.065] decile=3 0.046 0.107 [0.068] [0.055] decile=4-0.008 0.108 [0.052] [0.056] decile=5-0.081 0.237 [0.065] [0.049] decile=6 0.055 0.107 [0.066] [0.045] decile=7 0.013 0.121 [0.066] [0.048] decile=8-0.025 0.165 [0.065] [0.047] decile=9-0.045 0.156 [0.064] [0.043] decile=10 0.016 0.156 [0.062] [0.044] Observations 17829 17829 14057 14057 Source: LiTS. Robust standard errors clustered at the PSU level in brackets. p<0.01, p<0.05, p<0.10. Estimates from multinomial logistic regression; marginal effects reported. Only significant categorical regressors shown. Country fixed effects not shown. 24

Table A4. Observed mobility and perceptions of improvement in welfare No decile control Decile control Better off than parents No improvement in welfare Continuous improvement in welfare Better off than parents No improvement in welfare Continuous improvement in welfare Intergenerational mobility (education) 0.078 0.004-0.013 0.150 0.028-0.016 [0.028] [0.027] [0.013] [0.032] [0.036] [0.015] HH head -0.035-0.073 0.016-0.084-0.039 0.015 [0.036] [0.028] [0.010] [0.032] [0.024] [0.011] Male -0.001 0.078-0.008 0.026 0.033-0.009 [0.025] [0.024] [0.009] [0.031] [0.021] [0.009] Age: 25-34 0.060 0.048 0.022 0.089 0.027 0.006 [0.038] [0.032] [0.014] [0.044] [0.043] [0.014] Age: 35-44 0.114 0.125 0.002 0.133 0.085 0.003 [0.049] [0.048] [0.017] [0.048] [0.051] [0.016] Age: 45-54 0.077 0.220-0.049 0.101 0.169-0.035 [0.052] [0.041] [0.015] [0.050] [0.051] [0.018] Age: 55-64 0.128 0.370-0.090 0.131 0.309-0.075 [0.053] [0.047] [0.020] [0.053] [0.050] [0.019] Age: 65+ 0.169 0.472-0.112 0.238 0.366-0.099 [0.057] [0.043] [0.022] [0.060] [0.050] [0.022] Education: Secondary -0.048-0.017 0.012-0.101-0.076 0.017 [0.040] [0.027] [0.013] [0.043] [0.037] [0.014] Education: Tertiary 0.039-0.117 0.047-0.024-0.192 0.058 [0.052] [0.045] [0.018] [0.047] [0.058] [0.022] Unemployed -0.268 0.016-0.030-0.279 0.014-0.023 [0.040] [0.037] [0.020] [0.047] [0.042] [0.026] Urban -0.004 0.031 0.014-0.005 0.025 0.006 [0.027] [0.023] [0.011] [0.029] [0.026] [0.011] Home ownership 0.104-0.018 0.015 0.135-0.005 0.013 [0.035] [0.026] [0.012] [0.044] [0.026] [0.012] decile 0.013 0.005-0.001 [0.006] [0.005] [0.002] Observations 20865 18786 18786 15831 14631 14631 Source: LiTS. Note: Marginal effects from Probit estimation (evaluated at means). Country fixed effects not shown. Robust standard errors in parenthesis. p<0.01, p<0.05, p<0.1 25

Table A5. Correlates of perceptions regarding expectations relative to past and future generations Better off than parents No decile controls Next Better off than generation parents & same will be better for next off generation Better off than parents Decile controls Next Better off than generation parents & same will be for next better off generation Current (perceived) welfare ladder 0.050 0.030 0.030 0.053 0.020 0.028 [0.005] [0.005] [0.004] [0.006] [0.005] [0.004] Mobility in future only -0.089 0.007-0.020-0.081-0.003-0.011 [0.023] [0.024] [0.021] [0.027] [0.028] [0.023] Mobility relative to past only -0.087-0.076-0.051-0.085-0.063-0.047 [0.026] [0.026] [0.020] [0.030] [0.027] [0.021] No mobility -0.117-0.104-0.075-0.102-0.106-0.070 [0.023] [0.025] [0.019] [0.028] [0.025] [0.020] Round -0.050 0.071 0.009-0.038 0.062 0.020 [0.022] [0.022] [0.015] [0.025] [0.025] [0.016] HH head -0.029 0.018 0.003-0.035 0.019 0.008 [0.017] [0.016] [0.012] [0.017] [0.019] [0.014] Male 0.018-0.015-0.010 0.031-0.008-0.008 [0.013] [0.014] [0.010] [0.017] [0.015] [0.012] Age: 25-34 0.041 0.015 0.024 0.052 0.039 0.039 [0.026] [0.022] [0.020] [0.029] [0.025] [0.019] Age: 35-44 0.105 0.014 0.051 0.127 0.024 0.056 [0.030] [0.023] [0.023] [0.030] [0.026] [0.021] Age: 45-54 0.122 0.030 0.031 0.141 0.031 0.043 [0.030] [0.029] [0.022] [0.031] [0.028] [0.020] Age: 55-64 0.202-0.010 0.043 0.213 0.005 0.056 [0.029] [0.026] [0.022] [0.033] [0.031] [0.023] Age: 65+ 0.266 0.055 0.092 0.306 0.064 0.115 [0.032] [0.027] [0.023] [0.037] [0.031] [0.025] Education: Secondary -0.008-0.073-0.032-0.022-0.081-0.037 [0.018] [0.018] [0.011] [0.020] [0.020] [0.013] Education: Tertiary 0.044-0.065-0.018 0.026-0.057-0.016 [0.022] [0.020] [0.014] [0.021] [0.020] [0.014] Unemployed -0.166-0.045-0.075-0.155-0.023-0.056 [0.021] [0.019] [0.017] [0.025] [0.021] [0.019] Urban 0.010 0.010 0.005 0.001 0.014 0.003 [0.017] [0.019] [0.013] [0.020] [0.021] [0.014] Home ownership 0.073 0.017 0.025 0.070 0.027 0.038 [0.020] [0.018] [0.013] [0.022] [0.022] [0.015] decile 0.003 0.002-0.000 [0.003] [0.003] [0.002] Observations 33859 32667 32009 25214 24379 23964 Source: LiTS. Note: Marginal effects from Probit estimation (evaluated at means). Country fixed effects not shown. Robust standard errors in parenthesis. p<0.01, p<0.05, p<0.1 26

Table A6. Correlates of negative outlook on economy, financial situation, employment Worse national economic situation for next 12 mo. Worse HH financial situation for next 12 mo. Worse national employment situation for next 12 mo. Worse personal job situation for next 12 mo. Employee: low skill 0.007 0.028 0.004 0.009 [0.005] [0.004] [0.005] [0.003] Self-employed: high skill 0.005-0.016-0.016-0.000 [0.007] [0.006] [0.007] [0.004] Self-employed: low skill 0.000 0.043 0.021 0.017 [0.017] [0.012] [0.017] [0.009] Male -0.005-0.016 0.003 0.004 [0.005] [0.004] [0.005] [0.003] Age: 25-34 0.039 0.031 0.013 0.014 [0.009] [0.007] [0.009] [0.006] Age: 35-44 0.057 0.054 0.022 0.026 [0.009] [0.007] [0.009] [0.006] Age: 45-54 0.085 0.071 0.069 0.040 [0.009] [0.007] [0.009] [0.006] Age: 55-64 0.102 0.115 0.086 0.050 [0.011] [0.008] [0.011] [0.006] Age: 65+ 0.090 0.103 0.115 0.025 [0.018] [0.014] [0.018] [0.011] Secondary education or more -0.032-0.030-0.042-0.029 [0.005] [0.004] [0.005] [0.003] Household size -0.003 0.000 0.002 0.001 [0.002] [0.001] [0.002] [0.001] EU-13-0.019 0.131-0.072-0.042 [0.063] [0.045] [0.062] [0.047] year==2008 0.097 0.074-0.036-0.001 [0.006] [0.005] [0.006] [0.004] year==2010-0.007 0.016-0.036 0.018 [0.006] [0.005] [0.006] [0.004] year==2016-0.126-0.057-0.164-0.018 [0.006] [0.005] [0.006] [0.004] Observations 47030 47030 47030 47030 Source: Eurobarometer. Note: Marginal effects from Probit estimation (evaluated at means). Country fixed effects not shown. Robust standard errors in parenthesis. p<0.01, p<0.05, p<0.1 27